Hello! I recently got an alert that one of my dives was disabled. My server was relocated not long ago, so I'm hoping something simply got jostled loose. The connections appear okay, but I re-seated them anyway. My search of the forums has not brought up anything quite fitting my setup, and all links to unraid troubleshooting appear to be dead, I guess there were some pretty major changes to the website.

 

I have attached my diagnostics for the drive from before I rebooted. I haven't done anything since re-checking the connections, as I wasn't sure if there was something more I should do before attempting to rebuild the drive. Thanks for any insight!

ST4000DM004-2CV104_ZFN02B9J-20210901-1752 disk8 (sdi) - DISK_DSBL.txt

ST4000DM004-2CV104_ZFN02B9J-20210901-1752 disk8 (sdi) - DISK_DSBL.txt

Go to Tools - Diagnostics and attach the complete Diagnostics ZIP file to your NEXT post in this thread.

Done!

tower-diagnostics-20210901-1752.zip

3 minutes ago, ReverendDan said:

all links to unraid troubleshooting appear to be dead

You can access the latest version of the documentation by using the Documentation link at the bottom of this page, or by clicking the "manual" link in the bottom right corner of the webUI on your Unraid server.

Oh! Okay I'll check there.

SMART attributes for disabled disk 8 look OK, and emulated disk 8 is mounted so that all looks good.

 

Says you are currently running an extended SMART test on that disk. How long has it been running?

I thought the test had failed, as the progress indicator disappeared. I have since rebooted but will run the test again now

It will take several hours to do an extended test on 4TB. You should disable spindown on that disk or the test might get interrupted.

 

Looks like you were about 1:39 into your monthly parity check when disk8 started to get read errors, so Unraid tried to write the emulated data back to the disk to fix it and when that write failed it disabled the disk.

 

10 minutes ago, ReverendDan said:

I have since rebooted

Thank you for getting the diagnostics before reboot.

 

22 minutes ago, ReverendDan said:

My search of the forums has not brought up anything quite fitting my setup

I didn't notice anything unusual about your setup. Can you elaborate?

I just meant that others with this issue had different setups, mine is probably pretty standard.  The smart test ran for a while and the. Switched to aborted by host...

tower-smart-20210901-1946.zip

1 minute ago, ReverendDan said:

aborted by host...

 

Did you do this?

17 minutes ago, trurl said:

disable spindown on that disk or the test might get interrupted.

 

The disk is currently unassigned, I didn't see the option.  Let me do some digging, must have missed it 

Click on the disk to get to its settings page.

Unfortunately it appears there's no ability to adjust spindown on unassigned drives.  Some quick googling indicates that some people create a script to write a text file every 10 minutes to keep the drive awake...

Would it work to change the default for all drives?

9 minutes ago, ReverendDan said:

The disk is currently unassigned

Why did you unassign it? Did you try to mount it as an unassigned device?

I was following some instructions to remove and re-add the drive to have the parity rebuild the drive when I realized that I had done nothing to ensure the integrity of the drive. And rather than risk making things worse, I decided to stop and ask for help.  Better late than never, but I probably should have asked sooner.  Sorry to make things more difficult.  I appreciate the help!

Should I try to mount it?

Perhaps I should replace the drive with a larger one anyway.  I could run some preclears on this drive after the fact to test it.  Seems a decent excuse for an upgrade!

Just wondered if you had mounted it as an Unassigned Device. If you mount it read/write instead of read-only, it would be slightly out-of-sync with parity.

 

Ordinarily, a disk gets disabled when a write to it fails, and since the failed write updates parity so that write can be recovered, the disabled disk is out-of-sync.

 

But in this case it looks like a failed read during parity check caused a writeback to the disk, so it may not be out-of-sync or even need rebuilding.

 

Do you think there is any possibility that anything tried to write to the disk since the parity check began? If anything wrote to the emulated disk then the physical disk is out-of-sync and has to be rebuilt.

1 minute ago, ReverendDan said:

Perhaps I should replace the drive with a larger one anyway.

Rebuilding to a replacement is always the safest approach since it keeps the original in case there are any problems rebuilding. Often people won't want to buy a replacement though so they rebuild to the same disk.

 

2 minutes ago, ReverendDan said:

I could run some preclears on this drive after the fact to test it. 

You should definitely not preclear the original until you are satisfied with the replacement/rebuild.

 

Okay that sounds good.  If my local Costco doesn't have a larger drive I'll attempt to rebuild this one.  Is there any risk to the array if the rebuild fails?

There is always going to be some risk of hardware failure of course.

1 hour ago, trurl said:

SMART attributes for disabled disk 8 look OK, and emulated disk 8 is mounted so that all looks good.

But everything looks good for a successful rebuild. And you have dual parity.

 

Excellent.  Thank you for all your help!!